Morphologies of Comb-like Polyacrylic Acid/Polyacrylate Copolymers as Functions of the Degree of Derivatization with n-C₂₂H₄₅ Side Chains

Abstract

Polymers with crystallizable side chains have numerous applications, and their properties depend on their crystal morphologies and phase separation. Structural analysis on a wide spatial scale plays an important role in controlling the thermal properties and higher-order structures of these polymers. In this study, we elucidated the melting and crystallization processes of copolymers with varying crystallizable side-chain fractions over a wide spatial range. Differential scanning calorimetry revealed that the enthalpies of melting and crystallization increased linearly with increasing crystallizable side-chain fraction. The results of wide-angle X-ray scattering indicated that the crystal lattice was hexagonal. Conversely, spherulite-like higher-order architectures with linear structures and radial spreading were observed in the highly crystallizable components, but no micrometer-scale structures were observed in the less crystallizable components. In situ small-angle X-ray scattering was used to elucidate the phase separation and mixing processes. Lamellar crystallites were observed at crystallizable side-chain fractions of >55 wt.%, whereas small crystallites were observed at fractions of <45 wt.%. At temperatures above the order-disorder transition temperature, density fluctuations caused by correlation holes were observed. These properties have a strong effect on the crystallizable side-chain fraction.

1. Introduction

Comb-like polymers, which comprise polymer main chains and pendant side chains, exhibit characteristic crystalline structures, self-assembly morphologies, and melt properties [1,2,3,4]. Accordingly, their crystallization and segmental packing differ from those of typical linear polymers because of the influences of their structural compositions and side-chain lengths [5,6,7,8,9,10]. Consequently, the manipulation of comb-like polymers, which exhibit different behaviors in terms of side-chain crystallization, is useful in understanding their local chain-packing modes [6,11,12,13,14,15]. In addition, the polymer backbone, side-chain length, and chemical linkages significantly influence the characteristic chain packing and side-chain crystallization of a comb-like polymer. When longer side chains are introduced along the polymer backbone, an ordered packing structure is typically formed because of the increased number of crystallizable CH₂ units [16,17].

The attachment of appropriate side chains to their polymeric backbones may result in materials with remarkable structures and advanced functionalities, such as charge transport, in addition to optical properties and responsive performance [12,18]. Understanding the aggregated structures of comb-like polymers is advantageous in designing responsive materials with novel architectures. Furthermore, the pendant side chains do not fully participate in crystallization, and only the C atoms that are positioned away from the main chain enter the crystalline phase [4,16]. Thus, the minimum requirement or critical number of crystallizable C atoms varies with the stiffness of the polymer backbone [19,20]. In addition, the lengths of the side alkyl groups determine the crystal domains of the comb-like polymers, which exhibit distinct structural and packing patterns. Previous research regarding n-alkylated poly(p-benzamide) [21], polyethyleneimine [22], chitosan [23], and polyvinyl alcohol [24] comb-like polymers with side-chain lengths ranging from 12 to 22 carbon atoms showed that the melting temperatures (T_m) and enthalpies (ΔH_m) of the paraffin-like crystallites depend strongly on the sizes of the alkyl domains.

The comb-like polymer used in this study exhibits the same chemical structure as Intellimer^TM, which is an acrylic adhesive [25,26]. Intellimer^TM is under development for use as a thermosensitive adhesive based on the significant changes in its various properties with temperature. It is also under consideration for use as a film whose transparency and opacity may be controlled with temperature. Its physical properties have been summarized in several studies [26,27]. In this study, we used a semi-crystalline graft copolymer of short- and long-chain esters of polyacrylic acid, with polyacrylic acid as the main-chain backbone. The long-chain ester, which behaves as a side chain, contains the linear alkyl group C_nH_{(2n + 1)}, which crystallizes at low temperatures. Heating to the melting point of the linear alkyl moiety causes a transition from a semi-crystalline to an amorphous state. This transition may modulate properties such as the thermosensitive viscosity, permeability, and adhesion.

Therefore, by manipulating the primary structures of the crystallizable side chains, it is possible to vary the temperature at which the phase transition between crystallization and melting occurs, i.e., to change the adhesion and release properties. Additionally, the melting point and viscoelastic properties may be altered by varying the side-chain density. Such adhesives find numerous applications in the field of electronics, including in lithography [28] and the fabrication of indium gallium zinc oxide panels [29,30,31]. In addition, as their viscoelasticities vary with temperature, they are used in microcapsule coatings for seeds, with an emphasis on sustainability [32,33], and in films to control permeability to air [34]. Additionally, graft polymerization of materials such as Intellimer^TM onto the surfaces of microparticles to generate temperature-dependent microparticles has been studied [35]. The relationship between the crystal structure of a comb polymer and its adhesive properties has been elucidated via detailed analysis of the crystal structure, of changes in the higher-order crystal structure with temperature, and of the adhesion and release properties of the material [36]. In this study, thermal measurements and spatially extended structural analyses were used to evaluate the crystallization and melting processes of comb polymers with different weight percents of crystallizable side-chain components. The fraction dependences of the crystalline components are the focus of this study. In situ wide-angle X-ray scattering (WAXS) and synchrotron radiation were used to evaluate crystal growth and melting, and we evaluated their sub-micrometer structures using in situ small-angle X-ray scattering (SAXS). Depolarized light scattering and polarized light and atomic force microscopy (AFM) were used to study the micrometer-scale structure of the material. Accordingly, the degree of crystallinity, crystal size, density fluctuation, and higher-order structure as functions of temperature were elucidated and summarized.

2. Materials and Methods

2.1. Materials

Figure 1 shows a simplified chemical structure of the graft copolymer consisting of acrylic acid, methyl acrylate, and monomers randomly grafted with acrylic acid-linked alkyl side chains. The main chains, which consist of polyacrylic acid, poly(methyl acrylate), and 1-docosanol (C₂₂H₄₅OH)-linked polyacrylic acid, were supplied by Nitta, Osaka, Japan. n-Alkyl C₂₂H₄₅, as shown in Figure 1, crystallizes. The graft copolymer Intellimer^TM is a trademark of Landec (Menlo Park, CA, USA), which has a commercial licensing partnership with Nitta. The ratio of acrylic acid monomers to side-chain C₂₂H₄₅/CH₃/H was controlled by mass.

Table 1. Ratios (wt.%) of the monomers with side chains of C₂₂H₄₅/CH₃/H and ratio of the crystallizable side chain (mol%).

Sample	Monomer (wt.%) with Side-Chain C22H45/CH3/H	Crystallizable Side Chain (mol%)
0 wt.%	0/95/5	0
20 wt.%	20/75/5	5.3
30 wt.%	30/65/5	8.7
45 wt.%	45/50/5	15
55 wt.%	55/40/5	21
70 wt.%	70/25/5	34
90 wt.%	90/5/5	65
100 wt.%	100/0/0	100

shows the proportions of 1-docosanol acrylate, methyl acrylate, and acrylic acid employed. The weight-averaged molecular weight for all samples is about 500,000.

2.2. Measurements

Differential scanning calorimetry (DSC) was performed using a DSC Q200 (TA Instruments, New Castle, DE, USA), which was calibrated using the T_m values of In and cyclohexane standards. Samples of ~5 mg were heated in Al pans from 0 to 100 °C at a rate of 10 °C/min. The melting or crystallization temperature (T_m or T_x) is the onset temperature of endothermy or exothermy.

SAXS and WAXS were performed using the BL-6A beamline (Photon Factory, High Energy Accelerator Research Organization, Tsukuba, Japan) at a wavelength of 0.15 nm [37,38]. The respective camera lengths used in SAXS and WAXS were 931 and 243 mm, and the respective detectors used in SAXS and WAXS were PILATUS 1M and PILATUS 100K (DECTRIS, Baden, Switzerland). The ranges of the scattering vector q (=4πsinθ/λ, 2θ: scattering angle, λ: wavelengths used in SAXS and WAXS) were from 6.0 × 10⁻² to 2.5 and 8 to 19 nm⁻¹, and the thickness of each sample was approximately 200 μm. During in situ SAXS and WAXS, the sample film was sandwiched between two polyether ether ketone films and mounted on a Linkam 10002L hot stage (Linkam Scientific Instruments, Redhill, UK) for temperature control. FIT2D software (version 12.077, European Synchrotron Radiation Facility, Grenoble, France) was used for data processing, including contrast control of the 2D patterns and the creation of 1D profiles of the obtained 2D patterns.

AFM was performed to observe the internal structure on the sub-micrometer scale using an Agilent 5500 (Agilent Technologies, Santa Clara, CA, USA). The tapping mode was used to measure the topography, and the phase contrast was measured. The probes used were PPP-NCL probes (NANOSENSORS, Neuchatel, Switzerland) with respective spring constants and resonance frequencies of ~50 N/m and ~190 kHz, yielding a respective scan size and image resolution of 20 × 20 nm and 256 × 256. After the samples were bonded to Si substrates, Kapton films were applied to their surfaces.

Depolarized small-angle light scattering using a PP-1000 system (Otsuka Electronics, Osaka, Japan) was performed to observe the micrometer-scale higher-order structures. The camera length was 50 mm, and the light source was a He-Ne laser (λ = 654 nm). The range of the scattering vector q was 0.30–2.0 μm⁻¹, and the sample, with a thickness of ~40 μm, was sandwiched between two cover glasses. The temperature was controlled using the Linkam 10002 L hot stage [39]. Optical microscopy was performed using a VW-5000 microscope (Keyence, Osaka, Japan) equipped with high-speed charge-coupled device (CCD) camera attachments [40].

Raman spectra were measured in the 400–100 cm⁻¹ range using a LabRAM HR Evolution (HORIBA, Kyoto, Japan) equipped with a Syncerity CCD detector at an excitation wavelength of 532 nm. The respective spectral resolution, laser power, exposure time, and number of scans used in Raman spectroscopy were 0.5 cm⁻¹, 100 mW, 15 s, and 5. Temperature-dependent Raman spectroscopy was performed using a T95-HS temperature controller (Linkam Scientific Instruments) [41].

3. Results and Discussion

3.1. Thermodynamic Phenomena

Figure 2a shows the DSC thermograms obtained at cooling and heating rates of 10 °C/min. In the thermogram of the 0 wt.% sample, T_m and the crystallization temperature T_x are not observed; thus, no crystallization occurs in this sample. In contrast, T_x and T_m increase as the fraction of crystallizable side-chain components increases above 20 wt.%. Figure 2b shows that the glass transition temperature of the 0 wt.% sample is approximately 24.2 °C. Figure S1 and Table S1 shows DSC thermograms of 55 wt.% for the definition of T_m and T_x and the peak temperature and finishing temperature for melting and crystallization. The correlations between the content of the crystallizable side-chain component and T_m and T_x are shown in Figure 2c. Even at 45 wt.%, the T_m exceeds that of 1-docosane (C₂₂H₄₆) crystals (43.9 °C) [42]. The higher melting point may be due to the attachment of the crystallizable side chains to the main chain, which hinders molecular motion. The enthalpy (ΔH) and the crystallization enthalpy increase as the weight percent of the crystallizable side-chain component increases [43].

3.2. High-Ordered Structures at the Submicron and Micron Scales

Figure 3a,b shows the depolarized small-angle light-scattering profiles and polarized optical micrographs of the 45 and 70 wt.% samples at 25 °C. Below 45 wt.%, a very low scattering intensity was observed, indicating that the microscale structure forms slowly, whereas at 70 wt.%, a peak was observed at about q_p = 0.7 μm⁻¹. Figure 3c shows 10 μm-scaled spherulite-like structures under 70 wt. % conditions, as observed via AFM with linear structures. From the depolarized small-angle scattering pattern, an average spherulite size, R_s, was calculated from the peak position, q_p, using Equation (1) [44,45].

$$R_s=\frac{4.09}{q_p},$$

(1)

Here, q_p is the magnitude of the scattering vector at the peak position in the four-leaf clover patterns in the inset of Figure 3a. Therefore, the R_s was 5.8 μm. Figure 3d shows the temperature dependence of the total integrated scattering intensity during heating. Scattering is no longer observed upon melting. The temperature at which scattering completely disappears coincides with the melting point measured using DSC, indicating that these higher-order architectures display crystalline-derived oriented structures.

3.3. Crystal Lattice Determined via WAXS

Figure 4a–c shows the results of WAXS of copolymers with different side-chain fractions following heating of their crystalline states. In the profiles of all samples, we observed that the diffraction peak at q = 15.3 nm⁻¹ starts to decrease above T_m because the crystals are melting. Therefore, these peaks were considered with respect to the crystal structure of the crystallizable side-chain component.

Figure S2 shows the results of WAXS of the copolymers at 0 °C. We could find no crystalline peak for the crystallizable side-chain component of 0 wt.%, with only an amorphous halo observed; thus, the main chain may not crystallize above 0 °C. If the alkyl portion of the side chain exhibits the same orthorhombic structure as the polyethylene crystal, crystal-diffraction peaks should be present, in addition to the amorphous halo. However, only one amorphous halo and one diffraction peak were observed in this study. For a detailed analysis, we calculated the correlation length d based on the peak position using Equation (2).

$$d=\frac{2\pi }{q_{peak}},$$

(2)

where q_peak is the peak position in the scattering profile. Spacing values d_WAXS of 0.42 nm were observed in the hexagonal structures of the alkyl chains [11,46,47]. During heating, the diffraction peak representing the hexagonal crystalline structure decreases and shifts based on the smaller angle, and only an amorphous halo at q ≈ 14.0 nm⁻¹ is observed when the material is melted. We compared the crystallinities at different temperatures. The ratio of the crystalline regions in the WAXS profiles, R_c(T), is defined in Equation (3).

$$R_c\left(T\right)=\frac{I_{cry}\left(T\right)}{I_{cry}\left(T\right)+I_{amo}\left(T\right)},$$

(3)

where I_cry(T) and I_amo(T) are the respective areas of the crystalline and amorphous portions of the WAXS profiles. Figure S3 shows the decomposition of the 70 wt.% sample at 0 °C based on the WAXS profile. R_c(T) depends on the crystallinity, and we evaluated the crystalline size, L_hkl, using the Scherrer equation (Equation (4)).

$$L_{hkl}=\frac{k\lambda }{\beta \cos \theta }$$

(4)

Figure 5a shows the R_c(T) and L₁₁₀ of the side-chain crystalline component at 0 °C, which is below the melting point. R_c(T) increases with increasing fraction. In particular, in the case of 70 and 90 wt.%, R_c(T) increases slightly. These results are consistent with the results for ΔH from DSC thermograms in Figure 2c. Similarly, L₁₁₀ increases with increasing fraction.

Figure 5b,c shows the respective temperature dependences of R_c(T) and L₁₁₀ during heating. R_c(T) decreases with increasing temperature, becoming zero above the melting point determined via DSC. At fractions above 45 wt.%, L_hkl increases with increasing temperature, while R_c(T) decreases with temperature. These results suggest that the increase in L₁₁₀ is due to the aggregation and enlargement of small crystals with partially melting side-chain crystals. However, at a low fraction (30 wt. %), L₁₁₀ is independent of temperature, even at temperatures close to the T_m. Linear long-chain organic molecules are known to form lamellar intermediate phases (also called rotator phases) between their fully ordered crystalline phases and their isotropic liquid phases. The review by Cholakova and Denkov suggested that endothermic or exothermic peaks were observed during phase transition of the rotator phase of n-alkane [48]. Even in graft copolymers, endothermic peaks were observed [49]. We discuss the rotator phase between T_m and T_{m_finish}. From the report by Cholakova and Denkov [50], the difference between the melting point and the transition temperature of the rotator phase of C₂₂H₄₆ is 1.4 °C. On the other hand, the difference between T_m and T_{m_finish} is between 10.4 and 16.3 °C in Table S1. These data suggest that both the transition from fully crystalline to rotator phase and the transition from the rotator phase to isotropic melt occur over a wide temperature range in the case of comb-like copolymers, making it difficult to distinguish between these two translon processes. Consequently, one endothermic peak was observed in Figure 2a.

3.4. Sub-Micrometer-Scale Structure Determined via In Situ SAXS

Figure 6a,b shows the SAXS profiles obtained at 0 and 100 °C, respectively. As shown in Figure 6a, the peak at q ≈ 1.5 nm⁻¹ shifts to a higher q as the fraction of the crystalline alkyl side chain increases, and the profile of the 0 wt.% sample displays very weak peaks. Therefore, the peak represents the ratio of the polyacrylic acid main chains to crystalline alkyl side chains.

Meanwhile, as shown in Figure 6b, broad peaks were observed at q ≈ 1.5 nm⁻¹ and the secondary peaks are no longer observed. The correlation length d is calculated based on the peak position q_peak of the SAXS profile using Equation (1), and

Table 2. Correlation lengths of the samples with different weight percents of the crystallizable side-chain component annealed at 0 and 100 °C.

Sample	Correlation Length at 0 °C [nm]	Correlation Length at 100 °C [nm]
20 wt.%	5.92 ± 0.05	6.16 ± 0.07
30 wt.%	5.58 ± 0.07	5.44 ± 0.08
45 wt.%	5.08 ± 0.01	4.83 ± 0.08
55 wt.%	4.62 ± 0.02	4.34 ± 0.07
70 wt.%	4.09 ± 0.02	3.85 ± 0.07
90 wt.%	3.59 ± 0.02	3.33 ± 0.07

shows the correlation lengths d at 0 and 100 °C. Above 30 wt.%, the correlation length at 100 °C is smaller than that at 0 °C. The correlation length decreases with an increase in the content of the crystallizable side-chain component. Based on the DSC thermograms, the crystallinity increases as the content of the crystallizable side-chain component increases. The correlations calculated from SAXS measurements are those between crystals and crystallites or between crystalline and amorphous components. A low crystallizable side-chain component weight percentis associated with a low a correlation length between crystals. In other words, the increase in the correlation length with a decreasing crystalline component weight percent is due to the increase in the amorphous component, which increases the thickness of the amorphous region between crystals and thus the distance between crystals.

Figure 7a–c shows the temperature dependences of the SAXS profiles of the samples with crystalline alkane side-chain fractions of 30 (a), 55 (b), and 70 wt.% (c) during heating from 0 to 100 °C. There is a sharp peak at q = 1.5 nm⁻¹ due to the crystalline phase. As shown in Figure 7b,c, a secondary peak is observed close to q = 3.0 nm⁻¹ at temperatures below that at which the crystals are completely melted, T_{m_finish} in Table S1. Figure 6a shows the same secondary peaks in the profiles of the 45, 55, 70, and 90 wt.% samples. These results suggest that these copolymers contain lamellar microdomains with sizes of 3.59–4.62 nm. The correlation length is 1.3~1.6-fold longer than the fully extended crystallizable side-chain length of 2.8 nm. Thus, the SAXS profiles indicate that the polymer’s main chains form a lamellar plane with the n-docosane side chains. We evaluated the crystal thickness of the 100 wt.% sample using SAXS and DSC, and Figure S4 shows the Raman spectra obtained during heating. The longitudinal acoustic vibrations of the molecular alkyl chains were observed using low-frequency Raman spectroscopy.

Above T_{m_finish} or at a low crystalline alkane side-chain weight percent, only the broader peak is observed, with no secondary peak, i.e., when the secondary peak is absent, as shown in Figure 7a, the observation of one SAXS peak may be insufficient to support the conclusion that a lamellar morphology has formed within the polymer complex [51,52,53,54,55,56,57,58]. Therefore, we observe small crystallites, even below T_m. At the T_m of the onset temperature of melting, the peaks of SAXS profiles of the samples becomes weak. At the temperatures at which the melting process finishes (T_{m_finish}; in Table S1), only single broad peaks are seen due to the disordered states of the polymers. Thus, the complexes may exhibit characteristic copolymer-like fraction fluctuations (correlation hole effects) [53,54,55]. For a more detailed analysis, we focused on the order-disorder transition (ODT) and the melting process at about T_m. The change in the scattering pattern across the ODT enables the determination of the ODT temperature (T_ODT). There are several methods based on mean-field theory of identifying T_ODT using the results of SAXS. Two of the most common methods used are plots of (i) the reciprocal of the maximum scattering intensity (I_m⁻¹ (q_m)) against the reciprocal of the absolute temperature (T⁻¹) and (ii) the correlation length d_SAXS (=2π/q_m) against T⁻¹. According to Leibler’s mean-field theory, I_m⁻¹ (q_m) should decrease linearly with T⁻¹ in the disordered state:

$$I_m^{-1}\left(q_m\right)~a-\frac{b}{T},$$

(5)

where a and b are positive constants [56,57,58]. Equation (5) assumes the temperature dependence of the Flory-Huggins segmental interaction parameter between the polyacrylic acid main chain and crystallizable alkane side chains. Leibler’s theory predicts that d_SAXS or q_m in the disordered state is essentially independent of the temperature if we omit the small temperature variation due to the temperature dependences of the radii of gyration of the copolymer chains R_g(T).

$$d_{SAXS}/R_g\left(T\right)~T^0.$$

(6)

Thus, in the context of mean-field theory, the deviations in the temperature dependences of I_m⁻¹ (q_m) in Equation (5) and d_SAXS in Equation (6) are attributed to the onset of the disordering process upon heating, and they may thus be used to determine T_ODT. Figure 8a–c show plots of I_m⁻¹ (q_m) and d for the different crystallized copolymer components as functions of T⁻¹. As shown in Figure 8a–c, I_m⁻¹ (q_m) and d_SAXS are constant with respect to T⁻¹ at temperatures < T_m because the crystals in small crystallites (Figure 8a) and the lamellar structure (Figure 8b,c) prevent an increase in density fluctuations during heating. The gray region in Figure 8 shows the temperature during melting process from DSC measurements in Figure 2a. A linear decrease in I_m⁻¹ (q_m) is observed at temperatures above the crystal melting temperature, which is a result consistent with the theoretical predictions of Equation (5). Based on Equation (6), d is independent of T (or T⁻¹) but increases slightly with T⁻¹, and Sakurai and Nandan also observed a slight increase in d [51,54]. This increase occurs because the copolymer may exhibit an upper critical solution temperature-type phase separation, and the components of the main chain and crystallizable side chains should be mixed at the molecular level at temperatures > T_m. At temperatures just above T_m, I_m⁻¹ (q_m) exhibits a peak, and d decreases abruptly when the temperature increases, as shown in Figure 8b,c. This behavior is similar to the temperature dependences of star polymers [59]. The respective T_ODT values of the 55 and 70 wt.% samples are approximately 69 and 80 °C. Conversely, based on Figure 8a, T_ODT is 56 °C, which exceeds the T_m values of the crystallites in the 30 wt.% sample.

3.5. Modeling the Large-Scale Structures of the Grafted Copolymers with Crystallizable Side-Chain Components

Based on these results, models of the crystallizable side-chain copolymers were developed. First, when the weight percent of the crystallizable side-chain component is low (below 30 wt.%), very few crystals were observed, and the morphology depends on the main chain, which is almost randomly coiled. Therefore, small crystallites with low crystallinities were observed. As the crystallizable side-chain weight percent increases, the crystallinities and widths of the crystals also increase. WAXS revealed crystals with hexagonal structures, with correlation values of 0.42 nm. SAXS revealed lamellar structures with sizes of several nanometers in the more crystallizable components, as the correlation length of the lamellar structure is sufficiently smaller than that of the C22 extended chain. Conversely, small crystallites were observed within the less crystallizable components. Above T_{m_finish}, the crystallites are completely molten and an ODT occurs. Figure 9 shows a schematic summary of the structure of the copolymer. However, when crystallizable side-chain components were present below 20 wt.%, we could not analyze the small-crystallites and phase separation because of very small crystallinity. When the weight percent of the crystallizable side-chain components is above 90 wt.%, the sample amount is very small, making it difficult to analyze the phase separation and crystal morphology for in-situ SAXS/WAXS measurements.

4. Conclusions

In situ observations were used to elucidate the melting and crystallization processes of alkyl side-chain polymers with crystallizable side-chain fractions over a broad range of spatial scales. As the weight percent of the crystallizable side-chain component increased, T_m and T_x increased. The ΔH values of the crystals were also evaluated. In addition, the results of WAXS revealed that the crystal lattice was hexagonal, and SAXS revealed a several-nm-scale lamellar structure with highly crystalline components, as the correlation length of the lamellar structure was sufficiently smaller than that of the C22 extended chain, based partially on Raman spectroscopy below T_m. Between T_m and T_{m_finish}, it is difficult to distinguish the processes of transition from crystalline to the rotator phase and melting by DSC. Above T_m, the crystallites were molten. Further, above T_{m_finish}, an ODT occurred. At a larger scale, a spherulite-like higher-order architecture with a linear structure and radial spreading was observed.